Pengembangan Metode Klasterisasi Data Berbasis Hybrid Improved Artificial Bee Colony (IABC) dan K – Harmonic Means

One of data grouping process method is k-harmonic clustering method (KHM) which has a relatively short and simple process. However, it has a weakness at cluster center point. Randomly formed cluster center point causes difficulty to converge solutions. One way to solve the problem at the cluster center point requires a method which has a global solution for KHM. The method is Improved artificial bee colony (IABC), improvement of artificial bee colony (ABC) method based on behavior patterns of honey bee colony in food searching process. Advantage of the IABC method is able to have more optimum global solution. This research proposes a new method of clustering using improved artificial bee colony and K-Harmonic means (IABC-KHM) to optimize the center point in clusters that lead to global solution. In this study, the IABC is functioned for finding the most optimum cluster center point for the data clustering process using KHM. Furthermore, the performance test of the IABC-KHM clustering method is compared with ABC and ABC-KHM methods on three different datasets. The result of mean value of best function of IABC-KHM method of Iris dataset is 152,87, Contraceptive Method Choice dataset is 918,54, and Wine dataset is 31,01. Moreover, the result of the average value of the best F-Measure method IABCKHM Iris dataset is 0.90, the Contraceptive Method Choice dataset is 0.41, the Wine dataset is 0.95. To conclude, IABC-KHM method has successfully optimized the position of cluster center point that directs the cluster result which has global solution.